Leading-edge vortex monitoring in dynamically stalled flows via persistent homology

• Persistent homology can be applied for vortex core identification in dynamically stalled flows. • Template matching can be accelerated via down sampling with small accuracy penalty. • The vortex template size is a highly sensitive parameter at small radii. • An adaptive interrogation region can be used to accurately track the vortex core through various stages of the dynamic stall process. A novel vortex core identification pipeline is developed based on template matching. Using persistent homology, a template similarity field is constructed from a sliding window template-target feature space distance. This scalar field is then used to accentuate localised regions of spanwise vorticity via nonlinear weighting. This method is successfully applied to track the leading-edge vortex trajectory in a stall flutter starting cycle for a pitching NACA 63(3)418 aerofoil. Trajectory results are compared with several user-based vortex core identifiers like local vorticity minimum, local Q-criterion maximum, local swirling strength maximum, and manual tracking. The results of this comparison are quite satisfactory as the developed method is capable of automatically monitoring the leading-edge vortex core through several critical stages of its lifecycle. The effects of template size and down sampling are also investigated with respect to the vortex core identification. It is found that a template radius of r = 0.04 c and down sampling factor M = 10 are sufficient for accurate vortex core monitoring in dynamically stalled flows. In general, this method acts primarily as a field-based filter that can be useful for isolating highly vortical regions like the leading-edge vortex core in stall flutter or dynamic stall scenarios.